Spatial-numerical associations survive spatial frequency equalization in an adult unimanual SNARC task

The origins of spatial-numerical associations (SNAs), such as the SNARC (Spatial-Numerical Association of Response Codes) effect, remain debated. The Brain's Asymmetric Frequency Tuning (BAFT) hypothesis recently proposed that non-symbolic SNAs may arise from hemispheric asymmetries in spatial frequency (SF) processing, rather than reflecting a genuine spatial representation of number. Here we examined whether SNAs persist even in adult human cognition when both SF cues and, crucially, bimanual response confounds are eliminated. Participants performed a Go/No-Go non-symbolic magnitude comparison task, in which lateralized dot-arrays were compared to a central reference. Critically, all stimuli were controlled for low-level visual properties, including luminance and the SF power spectrum, thereby neutralizing potential SF hemispheric asymmetries. In addition, responses were executed using a single central key, removing any contribution of bimanual spatial coding. Results revealed a robust ratio effect, consistent with Weberian numerosity processing. Crucially, a significant SNARC congruency effect emerged in both accuracy and reaction times: relatively smaller numerosities were processed more efficiently when presented in the left visual space, whereas larger numerosities showed a right visual space advantage. Converging evidence from continuous regression analyses further supported this spatial-numerical mapping, showing an asymmetrical advantage for smaller magnitudes in left visual space and larger magnitudes in right visual space. This pattern persisted despite lacking SF asymmetries and bimanual response mappings. Together, these findings weaken the BAFT hypothesis, but only in adulthood, by showing that non-symbolic SNAs cannot be reduced to low-level visual or response-related confounds. Instead, they support a higher-order representation of numerical magnitude likely mapped onto space.